Quantitative NMR analysis of active alkaloids in A. mexicana decoction

An NMR method was chosen to rapidly access information on the amount of active alkaloids in the traditional decoction. By this means, the three alkaloids allocryptopine, berberine and protopine could be quantified in the same spectrum. This

is possible because the NMR response can be made the same for all components, contrary to HPLC/UV or MS methods. In fact, the strength of the NMR signal is proportional to the number of nuclei, thus the number of molecules, present in the analyzed sample. Moreover, in quantitative NMR analysis, a universal reference standard can be used for the analysis of most materials. As a consequence, a quantitative determination of a specific compound does not require pure samples for calibration, which is of great interest for the determination of natural products not always available as standard compounds (Maniaraet al., 1998).

Preliminary analyses were conducted to find the best ¹H NMR conditions (solvent and temperature) for allocryptopine, protopine and berberine. Allocryptopine and protopine are alkaloids of the protopine-type, presenting bad resolution in the region of the methylene protons of the ten-membered ring, especially at room temperature. For this reason, the chosen signals were those corresponding to the methylene dioxide protons of each molecule.

A comparison of spectra for the three alkaloids in pyridine-d5at 42C is shown in Figure 3.20. Under this conditions, one of the methylene dioxide signals of protopine overlaps with that of allocryptopine (H5.95). Since protopine has a second methylene dioxide signal atH5.90, this one could be considered for quantification and its integral value was subtracted from that at_H5.95, for the quantification of allocryptopine.

The second step of the method development was to check the spectrum of the sample prepared from the traditional decoction. Two different extraction methods were analyzed: 1) liquid-solid extraction, using methanol and the freeze-dried decoction, and 2) CH2Cl2partition from the filtered decoction at room temperature. The first extraction method was the same used in the HPLC/UV determination of berberine previously cited (Diop, 2006). However, when sample was recovered in 0.6 mL pyridine-d5, solubility problems arose. On the other hand, the second method provided a spectrum where all signals could be observed at a signal/noise ratio higher than 20 (Figure 3.21).

The second extraction method was then chosen for the ¹H NMR quantification of the alkaloids in the decoction.

Figure 3.20. Comparative ¹H NMR spectra of allocryptopine, berberine and protopine (500 MHz, 42C, pyridine-d5).

8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0

Chemical Shift (ppm)

5.905.95

8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0

Chemical Shift (ppm)

6.17

8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5

Chemical Shift (ppm)

5.95

allocryptopine

berberine

protopine

3.0

Figure 3.21.¹H NMR spectrum of an analytical sample obtained using the extraction method 2 for theA. mexicanadecoction (500 MHz, 70C, pyridine-d5).

The third step in the method development was to demonstrate the linearity and accuracy of the method through a calibration curve established for allocryptopine. The calibration curve was obtained from the ratio between the area under the signal selected for the compound and that for anthracene (internal standard) corrected by their MW ratio. A calibration curve for allocryptopine was obtained in the range of 0.6 – 5.5 mg/mL in triplicate. The values of integrals and calculated concentrations are presented in Table 3.5, and the calibration curve is presented in Figure 3.22. The R² value was as high as 0.9999.

7.0 6.5 6.0 5.5

Chemical Shift (ppm)

5.90

5.95

6.17

O-CH2-O Allocryptopine

+Protopine

O-CH2-O Protopine O-CH2-O

Berberine

Table 3.5.Integrals and calculated concentration values in triplicate for the calibration curve of

Figure 3.22.Calibration curve of allocryptopine.

In fact, calibration curves are not necessary in the quantitative NMR determination of compounds because the absolute integral under a peak is always proportional to the number of nuclei responsible for the signal. Thus, the area of a specific signal can be used to establish the concentration of a compound provided that the signal area of an internal standard is obtained at known concentrations.

The concentrations of allocryptopine, protopine and berberine were finally measured in two available clinical batches of A. mexicana. A CH2Cl2 partition was conducted in triplicate aliquots of 10.00 mL of decoction and the concentrations in the traditional preparation (mg/mL) were calculated from the absolute integral values of signal selected for each alkaloid and that of the internal standard. Table 3.6 presents the results expressed in terms of concentration (mg/mL) in tea as well as % (w/w) in the plant, which was calculated by using the initial mass of plant weighed and the total volume of decoction obtained.

The concentration of berberine was demonstrated to be similar between both batches when assessed by QNMR. The concentration of allocryptopine and protopine varied between both clinical batches. But the concentration of these protopine-type alkaloids together was higher than that of berberine in both clinical batches, which

R²= 0.9999

0.00 1.00 2.00 3.00 4.00 5.00 6.00

0.00 1.00 2.00 3.00 4.00 5.00 6.00

Added concentration (mg/mL)

Calculatedconcentration(mg/mL)

corroborates the hypothesis that they are the main alkaloids responsible for the activity, as discussed before.

Table 3.6.Calculated concentrations of allocryptopine, protopine and berberine in the decoction of clinical batches ofA. mexicanaleaves obtained by the QNMR method.

Batch Compound Concentration in tea (mg/mL)*

% in plant (w/w)* CV (%)

Allocryptopine 0.110.01 0.470.03 5.63

AM 8 Protopine 0.120.02 0.510.02 3.84

Berberine 0.110.01 0.480.03 5.99

Allocryptopine 0.180.01 1.100.07 6.57

AM 11 Protopine 0.060.00 0.390.01 3.86

Berberine 0.070.01 0.430.04 8.37

* AverageSD

The HPLC/UV and TLC methods previously developed for the quantification of berberine in the Mexican poppy decoction had demonstrated that the amount of this alkaloid in a non-clinical batch ofA. mexicanaleaves (batch AM 3) was 8.90.3 µg/mL in the decoction. In this specific case, the analyzed sample was a solution of the freeze-dried decoction in methanol at 20 mg/L (Diop, 2006). This result is lower than that found by QNMR in clinical batches AM 8 and AM 11.

One possible explanation is the difference in the extraction method. In the HPLC/UV and TLC methods, the decoction was freeze-dried and the resulting powder diluted in MeOH at 20 mg/mL and centrifuged at 13000 rpm prior to analysis. It is important to notice that a precipitate was formed and not considered in quantification.

When the same extraction method was applied to the QNMR method, the freeze-dried extract was not soluble at 20 mg/mL in the suitable solvent (pyridine-d5) and at 0.5 mg/mL the ¹H NMR signals of interest could not be detected at a good signal/noise ratio. It was then important to develop an extraction method to get an alkaloid-enriched fraction. This could be achieved by CH₂Cl₂partition directly from the decoction followed by centrifugation at 5000 rpm. The fact that the decoction was immediately cooled to room temperature and partitioned in the organic solvent, without a freeze-drying step, may have increased the recovery of alkaloids. Nevertheless, recovery data were

obtained neither for the previous method HPLC/UV and TLC methods nor in the QNMR method.

Since one of the specific goals related to A. mexicanainvestigation was to find a simple and low-cost semi-quantitative method for the quality control of an ITM in Mali, the semi-quantitative TLC method previously developed was applied to batches AM 8 and AM 11 and to an additional clinical batch AM 9, using the same extraction method as the one for the QNMR method. The results of different methods are critically compared and discussed in the next topic.